Former White House adviser praises Middle East's 'AI conviction' and calls for bolstered partnerships

Ben Buchanan also addresses concerns over China-owned DeepSeek

The UAE and Saudi Arabia, two of the Arab world's biggest economies, have invested heavily in AI. Victor Besa / The National

A former White House adviser on artificial intelligence has lauded Middle East governments' commitment to developing AI and said the US should bolster partnerships to assist with their ambitions, while also maintaining a global lead in the technology sector.

"The conviction of Middle East governments is certainly remarkable in AI and the ambition is also remarkable," said Ben Buchanan, who served under former president Joe Biden.

Mr Buchanan made the comments on Monday at Johns Hopkins University’s 2025 Emerging Technologies symposium in Washington, where he was asked about the UAE and Saudi Arabia's approach to AI.

The Arab world's biggest economies have invested heavily in AI. The UAE, in 2019, broke ground on the Mohamed bin Zayed University of Artificial Intelligence in Abu Dhabi, the world’s first higher-learning institution dedicated to AI.

Mr Buchanan warned that investment alone would not necessarily yield success and that co-operation and strategic alliances with other countries, such as the US, could make the difference.

"One of the foreign policy questions under the Biden administration and now under the Trump administration is going to be 'how do we structure the partnership in a way that keeps the Middle East with the United States?'" he said.

The UAE has developed strong partnerships with US companies such as Microsoft, and in turn the technology giant has pointed out that the country is "a hugely important hub" for its AI ambitions.

In March, the partnerships were on display during the visit of UAE officials to Washington, where AI, data centre infrastructures, energy and chip exports were the focus of talks with senior officials in the Trump administration.

US AI rivalry emerges with DeepSeek and China

Mr Buchanan also spoke at length about DeepSeek, the China-based AI company that has rattled the technology world, along with the US stock market, in recent months.

DeepSeek's large language model achieves parity with OpenAI models but claims to do so with far less computing power.

Some have suggested that, if DeepSeek's claims are true, the AI semiconductors and graphics processing units that have fuelled incredible wealth and innovation in the US are not especially important and could soon be rendered moot.

The company's debut in January caught many by surprise but not former White House officials like Mr Buchanan, who said he and others had started to pay attention to DeepSeek in 2023.

"The stuff they were putting out was pretty interesting, and we read some of their papers and we knew they were very talented, but we also knew they were compute constraint," he explained, noting that US AI chip export policies paved the way for DeepSeek to try to find efficiencies and algorithmic improvements.

He said although conventional wisdom was that DeepSeek's breakthrough's would spell trouble for the US, it is proof that Washington's strategy of exploiting its computing power is working.

"They’re still incredibly compute constrained, this remains a US advantage, and we should continue to double down on that advantage," he said, also expressing scepticism over DeepSeek's claims it spent less than $6 million to build its AI models – far less than the industry standard.

“They’re not accounting for all the money they spent on research up until the release of the model,” he said.

He also said US AI company Anthropic spent a similar amount of money to train its models, nullifying DeepSeek’s breakthrough to an extent.

Although some of the DeepSeek media buzz has died down, the company's AI assistant app is still going strong, routinely ranking in the top 10 productivity app downloads in Apple’s App Store.

The company keeps publishing new research and its LLM offerings are quickly finding their ways into daily life in China.

R1, DeepSeek’s open-source language model, is now available from Microsoft on the company’s Azure AI and Foundry development hub, while chip behemoth Nvidia has said R1 runs very well on its systems.

Despite all the intrigue behind DeepSeek, US-based OpenAI, Google and Meta still have more market share in the AI sector, along with plenty of funding to keep making improvements.

Mr Buchanan said success is no accident and it goes back to the US ability to attract talent from around the world.

"A lot of the people who are making great discoveries were not born in the United States, it’s our super power that we brought great people from all over the world to come and make contributions," he explained. “I hope we don’t lose that super power.”